LSTM_CNN Flashcards
Why does overfitting occur?
- The training data size is too small and does not contain enough data samples to accurately
represent all possible input data values. - The training data contains large amounts of irrelevant information, called noisy data.
- The model trains for too long on a single sample set of data.
- The model complexity is high, so it learns the noise within the training data
What is Overfitting?
Overfitting is an undesirable machine learning behavior that occurs when the machine learning
model gives accurate predictions for training data but not for new data.
How to avoid Overfitting?
One approach to reduce overfitting is to fit all possible different neural networks on the same
dataset and to average the predictions from each model. This is not feasible in practice, and can
be approximated using a small collection of different models, called an ensemble
What’s Dropout?
the practice of disregarding certain nodes in a layer at
random during training. A dropout is a regularization approach that prevents overfitting by
ensuring that no units are codependent with one another.
Question 1: What is overfitting in machine learning?
A) When a model performs well on the training data but poorly on new, unseen data.
B) When a model is under-trained on the training data.
C) When a model gives perfect predictions every time.
D) When a model is too simple to learn the underlying pattern of the data.
ANSWER: A
Question 2: Which techniques are used to prevent overfitting? (Select all that apply)
A) Reducing the number of layers in the neural network.
B) Employing dropout layers during training.
C) Using all data available for training.
D) Implementing data augmentation.
ANSWER: B & D
Question 3: What is the purpose of LSTM networks?
A) To classify images into different categories.
B) To improve the speed of training in neural networks.
C) To remember information for long periods of time in sequence prediction tasks.
D) To reduce the computational load during training.
ANSWER: C
Question 4: Which statements are true about the use of LSTM for stock price prediction? (Select all that apply)
A) LSTMs are unable to process time-series data.
B) LSTMs can capture temporal dependencies in stock price movements.
C) The model predicts stock prices based on the previous 60 days of prices.
D) High accuracy is not crucial because even moderate predictions can be profitable.
ANSWER: B, C & D
Question 5: What is Dropout in the context of neural networks?
A) Increasing the number of neurons in each layer to prevent loss of information.
B) Randomly dropping units during training to prevent co-adaptation of features.
C) A technique to speed up training by reducing the number of layers.
D) Adding noise to inputs to improve robustness.
ANSWER: B
Question 6: What are the key components of the CNN used for CIFAR-10 image classification? (Select all that apply)
A) LSTM layers to process image sequences.
B) Dropout layers to regularize the model.
C) Convolutional layers to extract features from images.
D) Batch normalization to standardize the inputs within the network.
ANSWER: B, C & D
Question 7: What role does batch normalization play in training neural networks?
A) It slows down training by adding additional computation.
B) It normalizes the output of a previous activation layer by subtracting the batch mean and dividing by the batch standard deviation.
C) It prevents the model from fitting to the data.
D) It increases the need for dropout.
ANSWER: B
Question 8: How is the training data prepared for the CNN model working with the CIFAR-10 dataset?
A) Pixel values are normalized to be between 0 and 1.
B) Images are resized to 32x32 pixels, the size required by CIFAR-10.
C) Labels are encoded using one-hot encoding.
D) Images are converted to grayscale to reduce complexity.
ANSWER: A & C
